Abstract. In order to solve the problems of time-varying, nonlinear and large inertia in the pneumatic control system of the intelligent valve positioner, the PID neural network control strategy is used to control the valve opening. Based on the analysis of the principle of the neural network controller, the network structure and the formulas of the PID neural network controller are determined to improve the system control accuracy and the dynamic performance. The simulation results show that the PID neural network has better time-varying, less adjustment time and greater accuracy than the traditional PID. The experimental results show that the settling time of the control strategy is less than 4s and the accuracy of the strategy is less than 1%.